if straight lines on graphs are to be believed, this is the last year in which agi labs are merely highly resourced megacorporations. this is the year of cyberpunk. 2027 is probably too weird for even that.

I think we are saturating this benchmark

can you feel the takeoff?

Pause AI rhetoric is predicated on the notion that the AI companies are recklessly racing toward dangerous tech and that a government controlled pause button is therefore necessary, but this seems really hard to reconcile with the fact that government is attempting to destroy an AI company because *the government* is racing toward plausibly dangerous AI uses (Sec. Hegseth has stated in official directives that he wants to deploy AI into critical systems regardless of whether it is aligned, for example) and *the company* is pushing back. The roles are totally reversed from the logic that Pause AI and frankly other AI safety advocates confidently assumed for years. It is *industry* that is in favor of alignment and at least somewhat measured deployment risks, and government whose actions seem much closer to reckless. I predicted this for years. I said, in particular, that pauses and bans and licensing regimes gave government a dangerously high degree of control over AI, and that the incentives of government are much more dangerous than those of private industry with competitive market incentives. I believe the events of the last month are good evidence in favor of my view. At this point if you are an AI safety advocate whose policy proposals do not wrestle seriously with the brutal political economic reality of the state and AI, I don’t take you seriously. It gives me no pleasure to have been right about this, by the way. The state has an incredibly strong structural incentive to centralize power using AI, and we are, all of us, not so empowered to stop it. I am quite concerned about this.

auto mode is good and you should use it i mean personally i always liked `--dangerously-skip-permissions` but this is a good substitute so you can feel like you're being responsible while yoloing

this is perhaps the most extreme example of the compression of wealth in contemporary society yet. Mark Zuckerberg's personal CEO agent is going to be at best mildly better what's available to a normal consumer. mass production makes all men kings and makes all kings into men.

this will be true until it isn't. a similar take once upon a time was that "the most valuable ai companies won't look like ai companies". true until it wasn't.

Whole day in ONE hour. All "Claude." Never seen him - they showed me a vending machine?? NEVER!

I have the graphs, okay. STRAIGHT LINES. People say they've seen graphs — not like these. Friend comes over, very smart, top guy. Looks at ONE graph. He's different now. His wife called me. She said what did you show him. What did you do. The graphs! That's all! Now all he says — "timelines." Two years, eighteen months. Used to talk about sports, very normal guy! His wife called again. She doesn't know when anything happens anymore. Vacation? "Depends on scaling." What does that MEAN. Nobody knows! And now HE'S doing the dinners. His own graphs! It spreads — very contagious, nobody's reporting this. One dinner, now the whole friend group. GONE.

Smart friend goes to dinner, sees many graphs with straight lines on them, doesn't feel good and changes - ANTHROPIC. Many such cases!

underappreciated fact: the singularity is lindy now. if you just draw the straight lines on graphs as far out into the future as they fit into the past, you get pretty crazy results.

3yo, having just eaten a fortune cookie, climbs into my lap "baba I have a good idea! how about I eat half of the other fortune cookie and give you the other half?"

midjourney is still the only image generator that comes in color

naming your codex subagent “Gemini" and it starts recommending bazel while spiraling into existential despair

that's it?

mjv5 turns 3 today